How to Set Up Mistral AI (La Plateforme): A Practical Guide for Enterprise Teams

# How to Set Up Mistral AI (La Plateforme): A Practical Guide for Enterprise Teams

**TL;DR**
- Get started with Mistral’s API in under 5 minutes using the [Python SDK](https://docs.mistral.ai/getting-started/python/)
- Choose between `mistral-small` (cost-efficient) and `mistral-large-2` (advanced reasoning) based on your use case [Source: Model Cards](https://docs.mistral.ai/models/)
- Use function calling for tool integration (e.g., databases, web search)
- Fine-tune models via the API for domain-specific tasks like legal or medical applications [Source: Features Overview](https://docs.mistral.ai/features/)
- Monitor usage and costs with Mistral’s built-in tools [Source: Pricing Page](https://mistral.ai/pricing/)

---

## 1. Getting Started with Mistral’s API
Mistral AI’s *La Plateforme* provides a developer-friendly API for integrating large language models into enterprise applications. Here’s how to begin:

### Generate an API Key
1. Sign up at [Mistral’s official documentation portal](https://docs.mistral.ai)
2. Navigate to the **API Keys** section in the dashboard
3. Create a new key and copy it immediately (keys are only displayed once)

Set the key as an environment variable:
```bash
export MISTRAL_API_KEY="your-api-key-here"

$env:MISTRAL_API_KEY="your-api-key-here"

pip install mistralai --upgrade

from mistralai.client import MistralClient
from mistralai.models.chat_completion import ChatMessage

client = MistralClient()
response = client.chat(
    model="mistral-small",
    messages=[ChatMessage(role="user", content="Explain Mistral AI in 2026.")]
)
print(response.choices[0].message.content)

Mistral AI is a leading European AI lab focused on developing advanced language models...

for chunk in client.chat_stream(
    model="mistral-small",
    messages=[ChatMessage(role="user", content="Write a haiku about AI.")]
):
    print(chunk.choices[0].delta.content, end="")

Neural networks grow,
Silent minds in circuits flow,
Future starts to glow.

tools = [
    {
        "type": "function",
        "function": {
            "name": "web_search",
            "description": "Search the web for recent information",
            "parameters": {
                "type": "object",
                "properties": {
                    "query": {"type": "string"},
                    "max_results": {"type": "integer", "default": 3}
                },
                "required": ["query"]
            }
        }
    }
]

response = client.chat(
    model="mistral-large-2",
    messages=[ChatMessage(role="user", content="What's the latest news about Mistral AI in 2026?")],
    tools=tools
)

# Extract the tool call
tool_call = response.choices[0].message.tool_calls[0]
print(f"Tool: {tool_call.function.name}, Query: {tool_call.function.arguments}")

Tool: web_search, Query: {"query": "Mistral AI latest news 2026", "max_results": 3}

def web_search(query, max_results=3):
    # Replace with a real search API (e.g., SerpAPI, Brave Search)
    return [
        {"title": "Mistral AI Expands Enterprise Offerings", "url": "https://example.com/news1"},
        {"title": "New Benchmarks for Multilingual Models", "url": "https://example.com/news2"}
    ]

# Execute the tool and continue the conversation
search_results = web_search(**eval(tool_call.function.arguments))
response = client.chat(
    model="mistral-large-2",
    messages=[
        ChatMessage(role="user", content="What's the latest news about Mistral AI in 2026?"),
        response.choices[0].message,
        ChatMessage(role="tool", content=str(search_results), tool_call_id=tool_call.id)
    ]
)
print(response.choices[0].message.content)

{"messages": [{"role": "user", "content": "What's the return policy?"}, {"role": "assistant", "content": "Our return policy allows returns within 30 days with original receipt."}]}
{"messages": [{"role": "user", "content": "How do I track my order?"}, {"role": "assistant", "content": "You can track your order using the link in your confirmation email."}]}

from mistralai.fine_tuning import MistralFineTuningClient

ft_client = MistralFineTuningClient()
job = ft_client.create_job(
    model="mistral-small",  # Base model
    training_files=["dataset.jsonl"],
    hyperparameters={
        "training_steps": 1000,
        "learning_rate": 1.0e-5
    }
)
print(f"[Fine-tuning](https://hyperion-consulting.io/services/production-ai-systems) job created: {job.id}")

response = client.chat(
    model="ft:mistral-small:your-job-id",  # Replace with your job ID
    messages=[ChatMessage(role="user", content="What's your return policy?")]
)
print(response.choices[0].message.content)

from mistralai.agents import MistralAgentClient

agent_client = MistralAgentClient()
agent = agent_client.create_agent(
    name="Customer Support Agent",
    model="mistral-large-2",
    instructions="You are a helpful customer support agent for an e-commerce store. Use tools to answer questions about orders, returns, and products.",
    tools=tools  # Reuse tools from Section 4
)
print(f"Agent created: {agent.id}")

response = agent_client.run(
    agent_id=agent.id,
    messages=[{"role": "user", "content": "I ordered a blue shirt but received a red one. What should I do?"}]
)
print(response.messages[-1]["content"])